anyone got any code they'd like to share?
I'm looking to find out whether the trend of incedent of abuse or self harm can be compared across a sample. The dates of these intervention vary from person to person but i record the week they took place in. So I can get a consitent sample to compare say week 1 with week 50 or month (lunar) 1 with month 12. What I'd like to see is if there is a downward or upward trend for individuals then to show whether the whole sample is predominantly upward or downward. I can get a linear trend for individuals but I'm not sure how "right" this is statistically speaking nor do I know how to show whether this fits to the group - unless say I do a mean deviation on the curves (?). Other suggestions are Chi-Square and Pearson's r both of which I can code (though how to choose the sample is still somewhat of a mystery - particualrly as it should be a random sample). Any way all thoughts or comments will be taken on board and very welcome. One thing I thought is that it might be nice to show a polynomal or logarithmic slope too depending on the r2 of the curve (i.e which is closest to 1) but I'm not sure how to calculate the Intecept or Slope of these without excel- something I'm tryig to avoid. So particular help in this area would be nice - unless this is total madness (in which case stop me now!).
Want the best answers? See FAQ181-2886
I'm looking to find out whether the trend of incedent of abuse or self harm can be compared across a sample. The dates of these intervention vary from person to person but i record the week they took place in. So I can get a consitent sample to compare say week 1 with week 50 or month (lunar) 1 with month 12. What I'd like to see is if there is a downward or upward trend for individuals then to show whether the whole sample is predominantly upward or downward. I can get a linear trend for individuals but I'm not sure how "right" this is statistically speaking nor do I know how to show whether this fits to the group - unless say I do a mean deviation on the curves (?). Other suggestions are Chi-Square and Pearson's r both of which I can code (though how to choose the sample is still somewhat of a mystery - particualrly as it should be a random sample). Any way all thoughts or comments will be taken on board and very welcome. One thing I thought is that it might be nice to show a polynomal or logarithmic slope too depending on the r2 of the curve (i.e which is closest to 1) but I'm not sure how to calculate the Intecept or Slope of these without excel- something I'm tryig to avoid. So particular help in this area would be nice - unless this is total madness (in which case stop me now!).
Want the best answers? See FAQ181-2886